there primary source document if you can get in there depending on your age, the actual papers of the president even those those of you can get them. >> in the books you have read recently? >> with buckling tramp political system? >> bouwer recently. you're interested in what the kids are reading but nothing that i can think of. >> would you currently reading? >> big triple. uninteresting story about the labor unrest and idaho the late part of the 19th

century when the governor was coming into his house to open the gates it was blown up by a bomb. there is a man speaking tomorrow that involve walter johnson that was one of the greatest pitchers of the earliest 20th century taking the been denied an auto and magnifies it across the united states at that time. it is pretty cool. . .

as people i talked at explained to me why date did anything they were doing and how do you know that nyc do that? at some point they did it because they had always done it that way are they had some kind of rule that was not based on any research and so i sort of one around campaigns with some degree of skepticism about a lot of the practices that were taking place and the way people were spending money and devoting time and resources and as they

learned about people starting in academia who were doing randomized controlled trials being adopted by people in the political world and learned more about all that data targeting, based on basically revolutionized campaigns in the last decade, there was a major generational shift in addition to all of the sort of new forms of research changing the way campaigns operated they have this kind of cultural tension between a lot of the old practices in the way campaigns operated in the new empirical movements. so you know i covered the 2000 campaign for the "boston globe" and was out talking with candidates about the sort of message death death going back back-and-forth back and forth and after 2009 i really burrowed down in this burgeoning world of people doing empirical research

and campaigns in academia and institutions in washington and elsewhere and was sort of condenser that is where the untold story of a lot of what is really going on in elections these days are taking place. >> host: interesting. what you want the reader of this book too must understand? when they read the book what is that message you want them to come away with? >> guest: i want them to understand that what we see on tv and in the newspaper every day of the campaign is very little of everything that is going on. you know, campaigns are far more than what is said in their ads are some of their ads and where the candidate goes and where it's running mate or spouse goes and i think with their spokesperson says on tv i think too much of her campaign coverage basically assumes that this is sort of the entirety of the campaign. must campaigns especially at the presidential level most of the people are saying they are doing

other things and that sort of is the thing we see. the candidate giving a speech, buying epson someone going on to tvs to explain the ads they do in many ways has not changed in decades. stuff that is taking place with people who don't go on tv whose names are almost entirely unknown the outside world has changed dramatically and is changing every year and sometimes every month and campaigns are getting a lot smarter about what they do and why they do it and how they do it and i think the general campaign coverage is struggling to keep up, explaining to the voters and the viewers what they are doing, why and what sort of thinking that is undergirding all of that. >> host: if you are talking about, i'm thinking what is called the victory lap you have this picture in your mind that there is the slab. can you give us some perspective in these campaigns now how many people are working on this

stuff? if you can tell us a typical campaign, how many people are working on these campaigns and how many people would be working on the sort of data analysis and all the things you are talking about? >> guest: on some level on smart campaigns data analysis is sort of informing everything they do. there is more and more stop on the campaign that is tested now either through rigorous experimental methods or per some sort of even if it's not randomized a lot of the experiments are some sort of discipline testing so the smart campaigns you know and the obama campaign is emblematic of this basically thinking of almost everything the campaign does this somehow formed by data but you know, you get down to the state-level campaigns and they definitely are having you know people who are voter file managers who are doing the data and targeting people who are doing the data. you get up to the presidential

campaign and one of the things, after they won the nomination built what they call the data science team and increasingly that sort of function is becoming a core function of the campaign. it used to be that there were lists for fundraising or voter list and you could buy them from vendors or consultants and now, you know, basically it's a core function of a modern campaign to have people especially on the voter side just crunching and processing data. >> host: if any of us were to go into the romney campaign are the obama campaign and we were to look around the headquarters how many people -- is there a lot of young staff? what does it look like? >> guest: chicago dozens of people doing voter data and date and fundraising data on line analytics and every state there

are jobs that are data jobs, voter file managers targeting directors. the obama campaign while thousands of people around the country hundreds of them are directly interacting with data every day. >> host: do you think one of the parties, the republicans are the democrats, is more adept at using this technique or they'll sort of at the same level? >> guest: i think at the moment the democrats are ahead to some leapfrogging cycle to cycle with different parties and one of the things that he comes clear and as i look back over where innovations took place it is one of the stories in this book that these campaigns are really persistent innovation. we build the company and at last six months or 18 months. you don't really know if it is going to be in advance and you are concerned about market share in one day. nobody really has the end

incentive to -- typically the person who is the chairman of the board of the campaign, the candidate, there are two types. someone who has done it before and thinks they know how to do it and maybe has been doing it over and over again and is perfectly happy keeping things the way they were or somebody who is never done before and there is no reason why somebody he wants to be a senator or a governor or a sheriff who doesn't know how to run a political campaign and the more technically advanced campaigns are getting less candidate should be expected to understand these things. and so, campaigns usually are not places that are built to innovate and are not well to learn from their failure so when i look back over presidential campaigns you see how re-election campaigns are exceptional as to institutions. you know they exist more or less for four years in some form or another. they are then able to plan a

budget over four years, stepping over four years. they often have primaries that distract them in the short term and for that reason, the bush 04 campaign was revolutionary and using some of these microtargeting tactics and applying them and in the obama campaign in 2012 if similarly a place where where exceptional innovation and statistical modeling is taking place and part of that is that you can set out a research agenda and in chicago there was postmortem committees that were set up in november of 2008 were people who were doing data and targeting campaigns came together and put together reports. almost four years ago, what they did right and what they did wrong in what they want to fix in the future. is really unusual in the world of candidate campaigns or even parties where there is not usually that much continuity to say here's the problem.

if we want more than a billion dollars to answered what can we take on and so, that culture, that institution building that has taken place around the obama campaign has given the democrats a major advantage. partly because of the long primary season and the huge staff that they had that so many people in that campaign at junior levels came into contact with sophisticated tools and analytics in a lot of those kids who were a field director and using fairly advanced voter data in 2008 went out and went to state parties are labor unions and talk to institutions on on the left and there is a collimation that is taking place that i think is a little more active on the democratic side right now than on the republican side. >> host: that's interesting so you would give about -- obama the advantage with this sort of

type attack may? >> guest: the romney campaign proved itself in a primary, to be very smart, to be disciplined and the number of their biggest successes, sort of how they went about winning iowa and without drawing too much attention to doing it for a long time and how they won subsequent states, all of this pretty small for evidence of their the ability to outbox santorum and some of the other and opponents opponents by using data. but, they are scrambling to sort of build everything against the obama campaign and anything approaching the ambition against the obama campaign in six months are obviously going to be smaller in four years. >> host: do you think that this campaign is going to rely more on messaging in the presidential campaign or on the

targeting technics? what is the next? i mean how much -- does the message still matter or can you just fine-tune it and targeted better and that will be the difference? how you put that together? >> guest: the message still matters in one of the things, this whole narrative takes place 10 to 15 years, has taken place in an era of real partisan polarization so one of the 2007 being this really formative moment i think in terms of innovation and politics for two reasons. one is a close election and so campaigns that might have once said 1%, 2% more than a vote if i this campaign started, like i could use one or 2% and so they are starting to be a real focus on techniques that can provide very small but measurable boosts. that is one thing that happened

in 2001 thing is we realize we were moving in this deadlock polarized era. the chief -- chief pollster for the bush campaign in 2000 started writing a memo while the supreme court was still taking up the case and in it he looked to polling data from 1984, something like 25% of americans were party switchers and by 2000 i think it was 7% and we are basically in that era now. we are a small percentage of the population that is actually persuadable between candidates of each party in a general election. most voters are predictable. th may not see pollsters are a divided magazine camping but there are far fewer voters who are actively moving between the two parties as there were previously and so when you get into that environment it's a lot

easier for campaigns to individualize where they can get benefits by focusing on turnout or registration of their supporters as opposed to merely trying to persuade. this year 6% from a percent or 9% that are persuadable obviously those campaigns are going to focus on them but we have a far better science now and the understanding what motivates people to vote and a lot of that informed by behavioral psychology research and so the science of mobilization and turnout has got much better. the science persuasion still, it's still pretty vague and so i do think that there has been a sort of reinvesting in a lot of localization techniques in part because we have learned in the last decade how they work so now you have these two separate things. you know once you get to somebody what you can and do to increase the likelihood of voting by 2% and now we have

better targeting technics through data to figure out who you talk to them about what. i don't think of it is necessarily message or targeting that good campaigns do targeting and analysis on the front end and that allows them to understand in a far more precise, clean play through our their turnout targets who they don't need to talk to until the contest and to their persuasion targets are and if you're narrowing the universe of the people you're trying to persuade you can make your messages a lot sharper. you can sort of focus your qualitative research and focus groups and polling and experimental testing to get more closely to the question of what, whether it's at 7%, talking to the 7% that are not persuadable and not trying to come up with messages speaking to a far broader side of the electorate and trying to persuade them all at once. >> host: a nail in the book you mention different technics that are used actually get

people to vote because i known this country there is such a low level many times of participation. could you explain some of these techniques that are used to actually get people voting? >> guest: yeah. the best ever demonstrated tool for getting people to turn out and vote was measured experiment in michigan in 2006 in the republican primary. this consultant and voter vendor named mark graham there have this sort of practice of exposing people if they didn't vote. and he organized an experiment with these two political scientist at yale, don green and allen grover who had been pioneers in using basically the same methods people use for drug trials to apply to campaign techniques, randomized controlled trials so contributing a pharmaceutical placebo to do different types of phonecalls and then go after the election to see who voted in who didn't. for the first time we were able

then to disentangle cause-and-effect in particular political tools. and in michigan, they found that remarkably effective was sending voters a copy, a letter before the election that said dear nancy, as you may know your history as a voter as a publicly available document on file by the board of elections. as you may know there's an election coming up in a few weeks. after the election we will send you and your neighbor a copy of your voting history and it had a line with the recipient vote history like you voted in the senate timer and did not vote in the school board election and then had a bunch of people selected with addresses on the same block. he got death threats when these went out.

tons of other sort of schooling from people who felt there was a laden but it turned out to be remarkably effective as a form of psychological social pressure. the expectation that people would have their friends and neighbors know not how they voted or which candidate they voted for but whether they are sick to go out. so this paper was published in was revolutionary because it basically was like 10 times more effective than anything that has never been used to moda by people to vote. no campaigns campaigns parties institutions wanted to put their name on anything like this because it really did seem like some form of blackmail and there was a real fear of backlash but over the next four or five years, a lot of institutions many of them on the left, academics working with unions, the dnc candidates, try to experiment with how to soften

the language and so now widespread craft of sending a letter out that they called the thank you treatment and it says something like dear nancy publicly available records show you voted and i want to thank you for being a voter. there's there is another election coming up and perhaps afterwards i can thank you again for participating. it's nowhere near as effective as the threat to expose people but behavioral psychologist who psychologists who work in this space, gets to the same mechanism which is that i think that voters have become so conditioned they have a secret ballot. many of them don't know that their vote history is publicly available. every campaign knows how often you vote in an election and don't realize that they are visible in a sense and that they can be monitored and so this idea of painting to people there

some form of surveillance that can take place and that people the people will judge you, you know it's clear that people want to vote and who will want to be seen as voters so as voters, these experimental research has shown time and again that they will do it. >> host: does anybody use that in the primary because i know that is the place where people are not showing up. is that a technique they could use? >> guest: round the people -- around the country people have but in this i haven't heard of people using that. is become so widespread in certain areas that people are now there is no element of surprise for voters that nudges them that way and wants voters our culture to the idea that they are being surveilled then there is only so much knew a listing can get in their

behavior the next time. there is sort of a confidence but one of the things that has taken place and that i write about in the book is behavioral psychology has informed so many aspects of human life in the last generation or so. certainly economics has been transformed by behavioral insight. politics and political science, once political scientist sort of rediscovered the use of experiments to actually go out in the real world and test techniques and campaigns on voters, they sort of naturally went to behave real science textbooks and said what are the things people are using to motivate individuals to recycle or to buy certain products or change their investing habits and can we translate these things to voting or registering? and so there still is i think a lot of stuff that has been demonstrated other than politics as having a sort of

psychological mechanism that people who work in politics thing can be translated into political behavior. and so it's all, almost all determined by motivation and not persuasion but i think they're slightly far more that comes out of that world the changes language in texture and a lot of the political communication we have. >> host: do you think most of it right now is more weighted towards motivating people to vote? is that what you are saying and less bickering out the science between the persuasion? >> guest: yeah, part of it is the way people design these experiments the easiest way to measure is how somebody voted. you go to the board of elections after an election and they have updated the voter file and it's yes or no if they voted. for suasion relies on polling people before and after to see if they change their minds and you have to rely on them being honest and they are self reporting of their choices and

so the other reasons, a lot of people who started doing this work in academia were using basically nonprofit dollars through their institutions than they couldn't do partisan work which made it very difficult to do. persuasion in a campaign but if you are doing non-candidate specific tests you can spend dollars out of your research budget on it. there is a big body of work and a lot of it informed by the behavioral psychology. if there is far less sort of i think persuasion and one thing i've written about this here is how the obama campaign is trying to adopt these experimental methods to measure the effect of their mail and on line ads and their tv ads and is more and more expensive but it's a way of trying to break out of relying solely on focus groups that try to ask voters to imagine in a sort of artificial setting what

would change their mind and instead introduce them to information in the real world where they don't believe they are basically being examined examined and see if it actually does change their mind. so that is one of the ways of using these experimental techniques to see what really does move for voters and get out of the sort of the artificial setting of focus. >> host: so on that note with the obama campaign since they have had four years ago time to test it and now they have four years later, is is there any innovative or different techniques they are using now any more than -- >> guest: the programs are really the thing and they have been used in small narrow ways by some institutions the afl-cio, emily's list, over the years and in certain states but using them in a presidential campaign as the obama campaign has this year is a major

advance. sort of the way that the people of tested messages historically is that you either have a poll that asks somebody who they support and tells them a bunch of information asked them again and you see if anybody moves are you ask them direct. if i told you that mitt romney had to pay taxes for x number of years would it make you more or less likely to vote for him? some people move or tell you that they would move. people telling you that they would move is a pretty sort of conjectural thing. i would not cognitively trust anybody who told you what they would do under a hypothetical scenario if they learned a permission. they may or may not already know and that is part of the polls that they asked people if you knew, you may or may not already know so in focus groups you bring somebody in and you have a

dozen people and you show them and add an asked them again if they would change their mind and now you are prompting somebody to change their mind and they are being forced to watch an ad that they might otherwise tune out. so using these experimented form programs at the obama campaign has introduces them into the real world so they randomly assigned the mail randomly to the electorate or randomly assign tv ads to certain markets and then because there polling across those markets they can see who moved based on which message or which type of bad or mail and then because they have all this data, thousands of individual data points on each voter, they can look at the attributes of the people who didn't vote, not just the people who said that they would move so you can statistically modeled a trace of somebody who moves when they see an ad about romney's taxes instead of just pulling people in looking at the type

who said they would move. and then they sort of have gotten it down to a cycle of a few weeks to develop tasks, randomly assign and distribute these ads and then measure the movement. at the end of that you it you realize okay the people who are moving on the task -- tax message let's say or conservative retirees okay now let's go with the ads to conservative retirees and that is a radical new way of transferring some of these experimental techniques to measure persuasion but also getting rid of all the sort of conjectural hypothetical message testing that has been used in focus groups and getting to something that actually gets at the behavior of changing their mind. >> host: interesting. how expensive is this? how much money do these campaigns to raise to do this at

the level that you are talking about? >> guest: it is expensive but i think campaigns that invest intelligently and it will tell you that it should pay for itself. so you know one of the benefits of migratory getting is that it helps you in the early process of sorting through the electorate and so if you can cut 10 or 20% out of your persuasion because you are smarter about who you should be talking to but 10 to 20% less than your male budget, 10 to 20% less of your phone budget if it helps you by your tv smarter you are there saving money or redirecting resources. so they are campaigns that i am microtargeting ads as something they add on to their male budget and in that case it's an expense and it may help them win votes or not. campaigns that build certain

type of analytics before getting into their earliest or tg projections should become smarter about how they spend all their other money because they are not spending persuasion messages to a already voted for them or who may not vote for them and they are sending the right messages to them so i message about abortion, 18% of women that we model as actually movable on abortion messages and then figure out the rest of them get talk to about. so i think that you know, you know it's expensive to analyze and certainly expensive what the obama campaign has done to build an in-house data analytic shop with many people working in private-sector jobs that pay good money but jim messina the campaign manager will tell you

that allow them to be smarter and get smarter so more data can get in and more models and projections and that it pays for itself. >> host: i think we are going to take a break now and we will be right back. >> host: i think of microtargeting. everybody has heard the phrase. can you just explain what is microtargeting exactly and how is it used in these campaigns? >> guest: sure. the biggest, there are two types of communications. there is mass media with the candidate talking and trying to get stories in the newspaper and

the paid media which for the last half-century has been tv and radio ads. those are very crudely targeting certain media markets or a certain time of day or on certain channels that you can do a lot more specificity than that. the other type of communication is voter contact, talking to individual voters where you know who you want to talk to any talk about specific messages or target get out the vote reminders to them. either you can put together what campaigns call a universe of your targets for something and because people are registered at a specific address, have a phone number so you know who you are talking to at the household level, and it's one of the big challenges of any campaign going back to under two years in the united states. figuring out if we are talking to people individually, who are we talking to and what are we talking to them about? we saw a little ferret -- very

little information on each voter. your name, your age in and your address and their gender and in some places their race and then you have some information that you could sort of extrapolated about based on where they live and their precincts. you knew someone's party registration but not everybody votes certainly throughout the 20th century a lot of people registered democrats were voting republican and vice versa but looking at the precinct, hundreds of people made some sort of prediction of what the political orientation of the area was like so campaigns, as sort of computing power increased through the middle of the 20th century, campaigns were using precincts as a sort of way of dividing the electorate. there's a basic trios at every

campaign wants to do. who are my people going to turn out and vote for me already because i'm not going to waste time talking to them. who are the people that will definitely support my opponent. ones i know who is in the middle starting to figure out what i'm talking to them about so precincts offered a very basic way of dividing a state or congressional district or if you are a democrat say okay people in strong democratic precincts, i want to turn them all out because 75% democratic tree sync even if i accidentally turn out, even turn out people and accidentally turn out people who vote against me but the margin is large enough that an aggregate will be turning out votes for myself and all the northern republican precincts and then when i get into the precincts in the middle that are 40%, 60% they are individually identifying each voter to know who they will vote for and what

issues they care about and that is where people set up phone banks. setting up a phone bank is an arduous process in the 50s and 60's because calling was expensive so you have to go into the area and set up a room with a phone but as soon as you did somebody would start calling voters off the voter rolls from the phonebook and say who you plan to vote for? one of the things in the united states that is different in other countries -- then you start to come up with a list of how individuals plan to vote and often a little bit of information. the same thing would happen when you send canvassers out. tried to identify voters d they are able to start sorting people out. who do i need to keep targeting with malar persuasion phonecalls.

so this is the way campaigns have always worked. what changed in the end of the last century we are starting to get data about people. it was well beyond what they knew about their area and their voter registration told you about themselves. it was not traditionally political information but a lot of it was collected by commercial marketers and especially the credit rating agency which really had an interest in accumulating as much possible information about your demographics and your habits and your lifestyle and what happened starting in the late 1990s was people in politics discovered they'd could -- what they knew about you politically and what they knew about your precinct and all the sources of information and use fiscal modeling techniques to basically run algorithms that could

predict how you would answer the question and who you would vote for so instead of in a given state, having to set up phone banks are sent out canvasser's to knock on the doors of the people who are in the precincts that are in the middle, you can use these modeling techniques to predict what each of those people would say. then you would do the same thing which is the people who were in the middle, persuadable, you would know at an individual level you wouldn't be writing off people based on their checker for your their registration and so this technique really took hold in 2004. the bush campaign was the first major campaign that sort of successfully designed and implemented a system where this type of microtargeting form of campaign, what it does is he a sort of tons of available individual information and automates the process to predicting how somebody is going

to. you are no longer neutralized by -- triaged by precincts and now i can take to people who live in neighboring apartments and without knocking on either their doors, predict who is likely to be with me and who is like he to still be undecided and then come up with a separate treatment for each of them. >> host: the need for all the canvassers and traditional door-to-door walking about has that subsided? do we not be that as much in his these campaigns class. >> guest: i think we do differently in campaigns like the obama campaign is smart about that. they did send people out to do i.d.s and those i.d.s came in to their statistical models and in 2008 the obama campaign was basically modeling every week which was to say that they would put out tens of thousands of calls and talking to voters

in that date state and all the data for commercial vendors were feeding into these dollars. every week they came it with a new projection of how e. guinta vigil would vote and whether they would vote him what issues they cared about. that allowed you to sort of guide canvassers to talk to people about particular issues and so, it's not, there is less of the need to do all of your identification work individually it helps to have volunteer activities and the more data points you have the better your predictions will be but then one of the ships with the obama campaign this year they are using their canvassers far more for persuasion work and not just for i.d.s. starkly campaigns usually use volunteers to do the work of going out and asking people who they are going to vote for and that puts together a male and mail the so of my volunteers are knocking on doors and find out 100 people believe in abortion,

they will make abortion brochures and put them in their mailboxes. now, you can do a lot of the predicting which people are most likely to be responsive to your message on abortion and then maybe you use your volunteers especially like the obama campaign, you have a circuit of people willing to help, and maybe you send them out to knock on doors or call 100 people and have a real conversation with them. one of the things that has come out of this body of experimental work is that meaningful volunteer person-to-person communication means something. it has an effect that pays call centers don't have, you know and it has callers have what they call chatting calls where they don't just go by for bait them

that are able to sort of improvise a little bit and sound more natural. people on the other end of the phone, it makes them more likely to turn out to vote after they have had that conversation so one of the themes running through this whole book is that this is all high-tech modern stuff that is made possible by massive data taking place by advanced modeling techniques and computer power. but, if there has been a real awakening of the human interpersonal dynamic in politics, think we have seen over the last decade the number of campaigns, the dean campaign and the bush campaign in 04 and the obama campaign in 08 that have had a real volunteer culture and i don't think those things are in conflict. so i think that campaigns are

recognizing that because they are targeting better and that because these person-to-person human interactions actually do make a difference, that they can use, they can direct the time of their volunteers more effectively. campaigns are integrating the two in a way that makes them sort of self-supporting and not in conflict with one another so i think that is what you will see on sidewalks and all of the obama field offices opening up, campaign using advanced techniques and lots of data that they have spent quite handsomely for but it's all in the service of figuring out who you get your volunteers to talk to and when and about what? that is one of the things that when i look back at the decade-long period i write about in the book it's heartening that i think it's precisely, everyone wants to see this moment in politics.

they know more about you and they know more about what moves you and i think in many ways what we are seeing is also seeing the value and we can now measure the value of human interaction. smart campaigns want to figure out how to allocate those. >> host: so personal touch still matters. >> guest: and we know it. i think he used you used to be there was acknowledging that personal touch is better but campaigns are reticent to trust their volunteers to talk to people and so i think there was a certain type -- the obama campaign is centralized as any of them but since right now they are able to measure the types of interactions of people i think campaigns are willing to concede a little but of the control and not assume that their message is always processing carry the television ads or mail that is

designed in headquarters but that there is a benefit to having people talking to people especially if you have done the difficult invisible work of telling which people to talk to and when and about what. i think campaigns are still weighing in but i look at the obama campaign in 08 were quite good at molding the left rain, right rain approach to campaigns than they were able to spend lots of money on advanced data and lots of volunteers and field staff and i think working through that is one of the things in this era most are figuring out how campaigns draw lines and can balance their need and desire for control, to control their message in control and centralized their strategy but also maximize the energy and the talent and the value of their supporters doing things on

their own for other voters. >> host: there is going to be the swing states where this election is going to be decided i would imagine all of the stuff you are talking about is going to be focused on in some of the states. can you remind us what those states the states are and is that true, that most of these techniques are going to be just targeted? >> guest: yeah, and i think more states and may move week to week. one thing to realize is that not all swing states are equal. or alike and i think we tend to think of them as interchangeable and tend to think of this at a polling as a bunch of states are between 46 and 49% in the campaigns are swing states, they are all sort of the same category. ..

>> what that the is strategically you do different things in each state. you go to different parts of the state. it may look out were the best of the campaign competes equally but i am sure they were thinking where they put good tv ad on the obama campaign one

little example there was the small concert u.s. and the columbus than the cleveland but it was the county seat because they had seen registration numbers were lagging in this area. from those goals that day aggregated they had him do a concert oriented to registration and it wasn't they sent in john legend to turn people out. that is happening everywhere. to all states we talk about

to in october by each competes differently based on what is basically takes every name of every individual person and they are the turnout targets of louis this persuade double where it goes. >> with everything you know, will there be one case state? if you are interested in? >> to us because they're new two presidential politics it is interesting to see the campaigns to don't have day

reflective said of tactics one-party user not equally established to write the presidential campaign. after spending one year you have the most granular familiarity with some of these statistical models but the empirical research of cause and effect i have less and less safe of predictions those who will have very big things but there is a lot in campaigning that that i write about people in

politics learned what works and what doesn't people are far more aware of limitations and knowledge and the day are significant. what happens inside elections that is the unknown and anybody that speaks of a two much confidence probably cannot be trusted. >> we hear about other tactics but what does that to world look like? can you paid to a picture? >> there really basic question when you buy the ad on broadcaster cable is uh

first campaign to do large-scale but they randomly assigned the tv ads in three weeks across the state. those who were called the eight heads one is a professor at the university of maryland. do we really know? trying to listen to the ball game is almost certain they the bounds of the media

market for what we're used to deal with. if it this measure rainwear the signal goes. that looks nothing of the media market. how many points into which market? people who are inside our outside also points are a crude measurement the 100 points is the average time a person nbc's your ad. there has been a time they

could match individual growing -- voting habits. but it turns out is an average but it distributes on the crazy curve when you think of your friends tv viewing habits some people will see it once. very few people see it five times but it misses the way the tv ads have been bought for a generation. now they & when you buy tv who is actually a expose? they can make much smarter and less crude decisions because of that. with fewer people are watching tv.

live television period. the natural substitute is advertising because you can still run video. the medium cannot be replaced just by phone you can do a lot with motion-picture and also online. the big challenge now is knowing about how much people are online as offline. politics take place offline and people vote at their home address. republicans vote in republican primaries. we know the individual house but to know who they are and what district when they're

watching videos. the challenge now is bringing geographical taka being more the banner ads and until this year it was prairies a fund-raising outlet most of the reason was to measure what works with fund-raising on line. you can run the web ad to measure your response rate to figure who is donating and who is not. now to use as a persuasion tool broker that will allow campaigns to have a better

idea who they should talk to online more conventional tv. >> host: i have two more questions. should the voter be concerned the campaign's get more data? >> one thain we have always had to remember long before the internet came along but to get a call during dinner who knew your name and party registration saying they call the last week to talk to about abortion i'm calling to remind you can we send to a van to your house next tuesday night they were annoying but nobody found them invasive.

fundamentally what campaign to doing with more data is beyond that. your record is public whether the campaign collects them permission. at least on the internet because they are afraid of the backlash in to get put in with the consumer marketers that they bend over backwards to respect privacy restrictions with cookies online. they're not required to. on the subside they know more about you but they use it to talk to you about the things they want to talk about. they have no interest in

doing anything other than catering to your individual needs interest. >> host: we have one minute. what would you want to say to people listening listening, learning about this. what advice can you give to the voter for those in the swing? >> thien is that you don't think of are the places that are the biggest innovations, i canvassing scripps. if they now gone your door to ask a review questions, that is where the greatest insights have been brought to bear.